inception-project / external-recommender-dkpro-tc Goto Github PK

• Change request to newest version
• Remove base64 encoding/decoding
• Fix tests

Add sentence based classification

• Get example request
• Implement it

• training of only one model at a time
• training should be async and return immediately

@reckart @rentier

I tested around with POS and NER predictions and this issue of missing annotation came up again.

POS:
If I just ignore not annotated words I have no problems. If I consider them and assign them a label for not labeled yet, the amount of not annotated token will outweigh the actual annotated data, i.e. all predictions will be the dummy-label for its extremely high frequency weight.

NER:
If I do not annotate words without annotation each word will receive a NER-tag. The notion of "no-tag" is missing in the training data.

Thus, I think we need a flag to indicate what to do with not annotated data. Depending on the task, ignoring them makes the predictions useless.

The DKPro Parent POM already sets up many things (compiler level, encoding, Maven plugins, code coverage profiles, etc etc). We could save a ton of settings by inheriting from it.

When I run

⇒ curl -H "Content-Type: application/json" --data @jsonPredictRequestV2small.txt http://localhost:30500/predict

then the server replies with

{"timestamp":1538040463710,"status":500,"error":"Internal Server Error","exception":"java.lang.NoClassDefFoundError","message":"org/dkpro/tc/features/ngram/base/LuceneFeatureExtractorBase","path":"/predict"}